Probabilistic syntactic analyser for information retrieval systems

State-of-the-art approaches to the problem of syntactic analysis of natural language are considered. This subject is one of the most complicated problems in the artificial intelligence domain. Probabilistic analysers are characterised by lower development laboriousness, do not need high linguistic training; they operation is qualitative enough, and in most cases it can be improved by additional training. In the probabilistic algorithms of syntactic analysis the tree method of making decisions, various statistic presentations and grammar models, heuristic optimisation methods, etc., are used. The conclusion is made that the combined approach uniting both formal grammar descriptions and probabilistic analysis mechanisms, seams to be most efficient.